A New and Better Way of Measuring the Tone of Media Coverage

The traditional way of measuring the tone of media coverage, which has not changed for over 30 years, is to use positive, neutral or negative sentiment. This is simple, but it reveals little about the media coverage and the impact it is likely to have on an audience. Take any crisis, for example, where the vast majority of media coverage is going to be negative and knowing whether this is 96%, 97% or 98% reveals absolutely nothing of any value for the PR and management team.

Metricomm takes a new and entirely different approach to measuring tone. As well as reflecting how today’s media and consumer environments have changed this also reveals much deeper insights into the sentiment of coverage along with the impact it is likely to have on an audience.

Instead of positive, negative and neutral sentiment, Metricomm analyses coverage in terms of its emotional content. This has some big benefits:

1. Human decision-making is driven by emotions. Understanding how media coverage is likely to affect an audience in emotional terms delivers valuable business insights
2. Breaking tone of coverage down by emotion reveals much richer understanding of media content and its likely impact on business outcomes
3. For example, coverage that would be classed as negative frequently involves powerful emotional drivers, such as fear or sadness. Knowing this leads PR teams to make better decisions in terms of response and reaction to coverage
4. Understanding emotional content enables PR activity to be more effective, including the use of negative emotions to drive positive outcomes. For instance, our analysis of BBC Studios’ The Blue Planet television series revealed how negative emotions of disgust and anger in media coverage generated by the programme resulted in very positive outcomes in terms of consumer response to plastic

Metricomm analyses tone of media coverage using six basic emotions originally defined by the psychologist, Paul Ekman. These are: happiness, surprise, sadness, fear, anger and disgust.

The best way to demonstrate the effectiveness of this approach is by using an example. Figures 1 to 3 show a breakdown of emotion for media coverage across three different sectors – insurance, universities and hotels – each of which has been affected in different ways by the coronavirus pandemic.

Fear is the strongest emotion for media coverage across the insurance sector. Although not too surprising given obvious concerns around the pandemic, this analysis reveals that the fear emotion also plays strongly in media coverage that excludes Covid. The reason is down to the fear factor prevalent in a lot of insurance sector media coverage regardless of whether it is about Covid or not.

“OK, but that’s obvious given insurance is largely about risk” I hear you say to yourself, but is it really that obvious? All too often such coverage is pronounced as negative in traditional sentiment analysis and only by using emotion does this critical nuance become quite so evident.

What might also be surprising are the levels of happiness in insurance coverage, especially given these are higher for Covid versus non-Covid content. After all, how can anything related to coronavirus be happy? In fact, emotion analysis reveals media coverage to be rather more balanced than we might believe.

Much of this misperception is down to our brains being hard-wired to recall negative events more vividly than positive ones, driven by the potential danger these would have represented over the thousands of years of human evolution. For the record much of the happiness in Covid-related coverage for the insurance sector has been down to the more recent evolutionary-driven factors of reduced costs and refunds, especially for car insurance. Hardly something that our ancestors would have had to worry about!

Figure 2 shows emotion of media coverage for UK universities, with one of the main differences compared to the insurance sector being fear. This is much lower for non-Covid coverage for universities than with insurance, for the reasons already discussed.

Anger, however, is much greater for non-Covid coverage for universities, which largely reflects the issues bedevilling our higher education establishments around free speech and sexual assaults. The real danger here is that anger will turn to disgust unless the universities affected get a grip on what are always going to be divisive and highly-charged emotions; and, as we will cover shortly disgust is by far the most damaging emotion to be associated with.

Again, as with insurance, happiness is strong across both Covid and non-Covid related media content, the former being driven largely by encouraging news around UK universities’ world-leading work on a coronavirus vaccine.

A sector where coronavirus has been particularly damaging, of course, is hotels. Figure 3 reveals much lower levels of happiness across the board for this sector while fear associated with Covid-related coverage is even greater than for insurance.

Much of this derives from highly damaging news around the financial impact of Covid on both leisure and business travel, leading in turn to high levels of furlough and many redundancies by hotel groups. Another emotion this also increases is sadness given the impact financial hardship inevitably has on many people’s live

What the emotion analysis also reveals about hotels is a level of disgust, which as we have already seen above is by far the most damaging of all the emotions. An emotion felt by humans at the most basic level, disgust developed through evolution over thousands if not millions of years to ensure we repel anything of potential harm, ranging from rotting food to dangerous situations. This is another legacy from our ancient ancestors, whose experiences remain very much with us in the modern world through our emotion-driven brains.

Disgust in media coverage is to be avoided at all costs. However, for hotel brands this can be especially difficult given that many tragic and appalling incidents take place on their premises, from suicides to murders and rape. Although much of this might be unfortunate there is no escaping the fact that any association with such content is highly likely to have a very damaging impact on a brand’s reputation; and while lumping it under ‘negative’ might be convenient this does nothing to highlight a potentially very significant problem.

In summary, using emotion to analyse the tone of media coverage delivers infinitely better insights than positive/neutral/negative. We have seen from the sectors covered here, at least, that media focus of media just as much on issues of concern as on anger, a vital nuance that would largely be missed by traditional sentiment analysis. The implications of this for PR planning and management strategy are both obvious and clear.

In a world increasingly dominated by the emotion of pandemic, racial issues and care for the planet, the need for greater understanding of what drives our feelings and behaviour has never been greater.